In the past we have accepted codes to be used as "language" codes which were non-existent and have had as a result that we are not in compliance with the rules of accepted use for the ISO-639 codes. When codes for new languages are used that are not consistent with the existing ISO-639 codes (all two and three character codes) a language should not be accepted at all.
We can handle ISO-639 codes, it's no problem to assign the correct code if it exist. In my opinion, even languages without an ISO-639 code should be accepted; however, this shouldn't be controlled by us, but the decision should be made in the New language requests vote.
According to your rules, anyone (with help of a few friends or of a few sockpuppets) can re-open the Zorglub language (oldbies will understand which language is concerned). So, your proposal needs to mention the issue of constructed languages.
The policy is just a proposal at the moment; however, we'll take this into account. There is a Quenya language test in Incubator; I suppose we should delete it?
Besides, I see you wrote "The Foundation will also have to approve the domain". Errrrrrrr. I'd prefer we avoid such bottleneck. How about something like "if at least 20 votes with a very large majority", no approval needed. If less votes or less obvious support, then, the Foundation or the spc must approve before creation ?
I removed the approval part, I think it remained there from the time when it was imported on Meta and reworded by Daniel. As for your 20 votes suggestion - unfortunately there are often cases of flash voting, where 20-30 voters can easily appear out of nowhere and declare their support. We can counter this somewhat by forcing all voters to have accounts, but there are still problems with that. I think this has yet to be decided somehow. We should make a percent range of approval, like with RfAs on English Wikipedia; e.g. more than 75% support gets approved automatically, between 50% and 75% needs approval by Foundation/SPcom and less then 50% fails to get a wiki estalished.
Michal Zlatkovsky ([[incubator:User:Timichal]])
timichal wrote:
In the past we have accepted codes to be used as "language" codes which were non-existent and have had as a result that we are not in compliance with the rules of accepted use for the ISO-639 codes. When codes for new languages are used that are not consistent with the existing ISO-639 codes (all two and three character codes) a language should not be accepted at all.
We can handle ISO-639 codes, it's no problem to assign the correct code if it exist. In my opinion, even languages without an ISO-639 code should be accepted; however, this shouldn't be controlled by us, but the decision should be made in the New language requests vote.
You can only vote on things that are acceptable in the first place. It is not acceptable to make up codes that are in violation with the rules of usage of the ISO-639 codes. You can vote all you like and have whatever majority but is is fundamentally wrong. When you want to have codes that are permissible, the http://simple.wikipedia.org is a great example of what CAN be done without infringing on compatibility of the ISO-639 rules.
According to your rules, anyone (with help of a few friends or of a few sockpuppets) can re-open the Zorglub language (oldbies will understand which language is concerned). So, your proposal needs to mention the issue of constructed languages.
The policy is just a proposal at the moment; however, we'll take this into account. There is a Quenya language test in Incubator; I suppose we should delete it?
Besides, I see you wrote "The Foundation will also have to approve the domain". Errrrrrrr. I'd prefer we avoid such bottleneck. How about something like "if at least 20 votes with a very large majority", no approval needed. If less votes or less obvious support, then, the Foundation or the spc must approve before creation ?
I removed the approval part, I think it remained there from the time when it was imported on Meta and reworded by Daniel. As for your 20 votes suggestion - unfortunately there are often cases of flash voting, where 20-30 voters can easily appear out of nowhere and declare their support. We can counter this somewhat by forcing all voters to have accounts, but there are still problems with that. I think this has yet to be decided somehow. We should make a percent range of approval, like with RfAs on English Wikipedia; e.g. more than 75% support gets approved automatically, between 50% and 75% needs approval by Foundation/SPcom and less then 50% fails to get a wiki estalished.
The problem with the votes that were taken on Meta proved that there was little merit in much of the votes. Many of the votes were plain political. The lack of reason proved to be such that I only voted when the proposal was seriously flawed. For me ANY proposal that prevents us from communicating what language we are using is seriously flawed. It did not prevent us from having projects where an existing Wikipedia prevents setting up a wikipedia using the rightful "owner" of the ISO-639-3 code .. :(
Does it help that with single login, everybody who works on any project has his or her user ??
Thanks, GerardM
timichal wrote:
In the past we have accepted codes to be used as "language" codes which were non-existent and have had as a result that we are not in compliance with the rules of accepted use for the ISO-639 codes. When codes for new languages are used that are not consistent with the existing ISO-639 codes (all two and three character codes) a language should not be accepted at all.
We can handle ISO-639 codes, it's no problem to assign the correct code if it exist. In my opinion, even languages without an ISO-639 code should be accepted; however, this shouldn't be controlled by us, but the decision should be made in the New language requests vote.
I wouldn't go so far as to say a language without a code should be completely unacceptable, but the threshhold should be higher. Arbitrarily assigning an unused code to the language could be problematic at some time in the future. There is a block of "q" codes reserved in the standard for user use. That's probably what we should use if necessary.
According to your rules, anyone (with help of a few friends or of a few sockpuppets) can re-open the Zorglub language (oldbies will understand which language is concerned). So, your proposal needs to mention the issue of constructed languages.
The policy is just a proposal at the moment; however, we'll take this into account. There is a Quenya language test in Incubator; I suppose we should delete it?
Quenya is one of the 24 constructed languages with an ISO 639-3 code. Before a project is started in them the support should be overwhelming.
Ec
wikimedia-l@lists.wikimedia.org